GraphCodeBERT is an open-source pretrained model for programming-language understanding, based on RoBERTa and trained with code and data-flow information. Developers can download and run it with machine-learning frameworks such as Transformers, PyTorch, TensorFlow, or JAX.
In the Other models space, Graphcodebert Base takes a focused approach. It focuses on understanding and generating useful representations of source code and natural-language programming descriptions. It is built as an open-source project for machine learning engineers and developers. The project is open source (Open Source). It ships for the web, API, and the command line, and it can be self-hosted.
Behind Graphcodebert Base is Microsoft. Key capabilities include code understanding, data-flow modeling, and masked language modeling.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do